{"id":"W4409918354","doi":"10.1145/3733620.3733635","title":"Reservoir Sampling over Joins","year":2025,"lang":"en","type":"article","venue":"ACM SIGMOD Record","topic":"Reservoir Engineering and Simulation Methods","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Joins; Sampling (signal processing); Computer science; Programming language; Computer vision","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002568327,0.0006965027,0.001560284,0.0008602004,0.001150123,0.002114881,0.002427608,0.001020494,0.002418376],"category_scores_gemma":[0.01183294,0.0005465624,0.001003223,0.001788011,0.00137481,0.005193471,0.00292429,0.001560326,0.0005108699],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001089685,"about_ca_system_score_gemma":0.001812084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004282934,"about_ca_topic_score_gemma":0.004612875,"domain_scores_codex":[0.998052,0.000472048,0.0001281301,0.0004916806,0.0006492654,0.0002068727],"domain_scores_gemma":[0.9944293,0.003035513,0.0003819928,0.001234187,0.0006630878,0.0002558516],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005488785,0.0001619455,0.005466985,0.0002669075,0.0001343711,0.0002991309,0.0003274335,0.669062,0.01391635,0.1363299,0.007274947,0.1662111],"study_design_scores_gemma":[0.00001463763,0.00004048126,0.0001629353,0.00000699477,0.00001151344,0.00005975557,0.00003085972,0.9653066,0.003156757,0.02925965,0.001939744,0.00001009951],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01697906,0.0002860184,0.9799485,0.0001952823,0.00004985702,0.00007180731,0.0001620108,0.0009253531,0.001382079],"genre_scores_gemma":[0.485392,0.0005665555,0.5078584,0.0003089137,0.0001787711,0.0002844428,0.001135211,0.0003691037,0.003906569],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004282934,"threshold_uncertainty_score":0.01358283,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03661353942592897,"score_gpt":0.3200537669503436,"score_spread":0.2834402275244146,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}